BAPTISTA v. Legal Aid Justice Center; Capital Area Immigrants' Rights Coalition; American Immigration Lawyers Association; Maryland Office of the Public Defender; Massachusetts Committee for Public Counsel Services, Amici Supporting Petitioner. (2025)

Samir Fernandes Baptista is a lawful permanent resident who was convicted in Massachusetts of unarmed assault with intent to rob or steal. During Baptista's later removal proceedings, the immigration judge and Board of Immigration Appeals concluded that the state crime of which Baptista was convicted is categorically an aggravated felony attempted theft offense under the Immigration and Nationality Act. So the immigration judge ordered Baptista removed for having committed “an aggravated felony ․ after admission.” 8 U.S.C. § 1227(a)(2)(A)(iii). The Board upheld that order.

We consider whether the Board's legal conclusion is correct. Concluding that it is, we deny Baptista's petition for review.

I.

Baptista is a citizen of Cape Verde. He was admitted to the United States as a lawful permanent resident in 2005. More than a decade later, he was convicted in Massachusetts state court of unarmed assault with intent to rob or steal. He was sentenced to nine to ten years' imprisonment.

The Department of Homeland Security issued Baptista a notice to appear based on his conviction. The notice asserted that Baptista's state conviction was “an aggravated felony as defined in sections 101(a)(43)(G) and 101(a)(43)(U) of the [Immigration and Nationality] Act, a law relating to an attempted theft offense ․ for which the term of imprisonment is at least one year.” J.A. 479.

Baptista moved several times to terminate the removal proceedings. The immigration judge denied the motions, concluding that the state offense qualified as an aggravated felony attempted theft offense, which in turn rendered Baptista removable under 8 U.S.C. § 1227(a)(2)(A)(iii). Baptista appealed the immigration judge's ruling to the Board of Immigration Appeals, which “affirm[ed] the decision of the Immigration Judge,” and “dismissed” the appeal. J.A. 3–7.

Baptista timely petitioned for our review. We review the Board's legal conclusion de novo. Sanchez v. Sessions, 885 F.3d 782, 786 n.2 (4th Cir. 2018).

II.

A.

As he did before the Board, Baptista argues that his Massachusetts unarmed assault conviction doesn't qualify as an attempted theft offense under the Immigration and Nationality Act. He contends that a defendant may be convicted in Massachusetts of unarmed assault by stealing property with the victim's fraudulently obtained consent.

If Baptista is correct, then the offense would be broader than a generic federal theft offense, which requires the taking to be against the victim's will. He also presses that his conviction can't qualify as an attempt because it doesn't require a defendant to commit an overt act toward a taking.

As we explain, we reject Baptista's arguments.

B.

We start with a brief primer on the relevant law.

The Immigration and Nationality Act renders removable “[a]ny alien who is convicted of an aggravated felony at any time after admission.” 8 U.S.C. § 1227(a)(2)(A)(iii). “[A] theft offense ․ for which the term of imprisonment [is] at least one year” qualifies as an aggravated felony.¹ Id. § 1101(a)(43)(G). So does “an attempt ․ to commit” such a theft offense. Id. § 1101(a)(43)(U).

To decide whether a state offense qualifies as a theft offense under the Immigration and Nationality Act, we apply the categorical approach, under which “we consider only the elements of the statute of conviction rather than the defendant's conduct underlying the offense.” Omargharib v. Holder, 775 F.3d 192, 196 (4th Cir. 2014). We must identify the elements of the state offense, which are “the constituent parts of a crime's legal definition—the things the prosecution must prove to sustain a conviction.” Mathis v. United States, 579 U.S. 500, 504, 136 S.Ct. 2243, 195 L.Ed.2d 604 (2016) (quotations omitted). And those elements must be “the same as, or narrower than, those of the generic offense” for the offense to count as a categorical match under federal law. Id. “[I]f the crime of conviction covers any more conduct than the generic offense,” then it isn't a categorical match.² Id.

We deploy this analysis twice for attempt offenses, once for the offense that is the object of the attempt, and once for the attempt itself. United States v. McCollum, 885 F.3d 300, 304–05 (4th Cir. 2018). The object of the attempt—here, Massachusetts theft—must categorically match a generic federal offense.

The “generic definition of theft” is the “taking of property or an exercise of control over property without consent with the criminal intent to deprive the owner of rights and benefits of ownership, even if such deprivation is less than total or permanent.” Gonzales v. Duenas-Alvarez, 549 U.S. 183, 189, 127 S.Ct. 815, 166 L.Ed.2d 683 (2007) (quotation omitted); see also Mena v. Lynch, 820 F.3d 114, 117 (4th Cir. 2016) (explaining that lack of consent is essential). A theft under state law committed with the victim's consent doesn't fall within the federal generic definition of theft. Such a state crime therefore isn't a categorical match to federal theft.

Even if a state theft offense qualifies as a federal theft offense, an attempt to commit that state offense must still independently qualify as a federal attempt offense. That happens when a state crime's elements include “(1) an intent to commit a theft offense, and (2) an overt act constituting a substantial step towards the commission of the offense.” Rendon v. Holder, 764 F.3d 1077, 1084 (9th Cir. 2014) (cleaned up).

III.

We now turn to whether Baptista's conviction is a categorical match to an aggravated felony attempted theft offense under the Immigration and Nationality Act. Baptista was convicted of unarmed assault with intent to rob or steal under Massachusetts law.

The statute of conviction reads:

Whoever, not being armed with a dangerous weapon, assaults another with force and violence and with intent to rob or steal shall be punished by imprisonment in the state prison for not more than ten years.

Mass. Gen. Laws ch. 265, § 20. Accordingly, the elements of this offense are: (1) an assault (2) by an unarmed person with (3) force and violence and (4) the intent to rob or steal.

A.

We start with whether the completed offense contemplated by the statute of conviction—robbery or stealing—is a generic federal theft offense.

Baptista focuses on stealing. According to him, “stealing” under Massachusetts law includes a taking of property accomplished with the consent of its rightful owner (albeit fraudulently obtained), as in embezzlement or larceny by false pretenses. Since a generic federal theft offense requires the taking to be accomplished “without consent,” Gonzales, 549 U.S. at 189, 127 S.Ct. 815, Baptista contends that the theft contemplated by his statute of conviction is categorically overbroad.

If Baptista is correct, then his statute of conviction wouldn't be a categorical match to a federal attempted theft offense as to both the completed offense and the attempt itself. That's because the completed offense wouldn't be a generic federal theft crime and because the attempt defined in the statute wouldn't encompass the intent to “commit a theft offense.” Rendon, 764 F.3d at 1084.

As we explain, we reject Baptista's reading of Massachusetts law.

1.

First, some context.

At common law, larceny was “the unlawful taking and carrying away of someone else's tangible personal property with the intent to deprive the possessor of it permanently.” 50 Am. Jur. 2d Larceny § 1 (2025). Larceny was a “crime against possession” involving “trespass” against personal property. Id. § 9. In contrast, the common law distinguished other theft offenses—like larceny by false pretenses and embezzlement—where the owner consented to the taking based on fraud.

The modern trend is to abandon the distinction between “various common-law categories of crime, and instead group them under the catch-all term ‘theft.’ ” Id. § 2. Massachusetts has adopted this approach. Its larceny statute reads, “[w]hoever steals, or with intent to defraud obtains by a false pretence, or whoever unlawfully, and with intent to steal or embezzle, converts ․ the property of another ․ shall be guilty of larceny.” Mass. Gen. Laws ch. 266, § 30(1).

Massachusetts's highest court has explained that “three formerly separate crimes”—larceny, embezzlement, and obtaining by false pretenses—“have been merged into the one crime of larceny ․ to eliminate the possibility that a defendant indicted for one of the crimes would escape punishment if the proof at trial established another of the crimes.” Commonwealth v. Mills, 436 Mass. 387, 764 N.E.2d 854, 860 (2002). Accordingly, “the word steal has become a term of art and includes the criminal taking or conversion by way either of larceny, embezzlement or obtaining by false pretences.” Id. (cleaned up). And Massachusetts allows indictments to allege broadly that a defendant “stole” without requiring prosecutors to specify whether the stealing occurred through “larceny of the property,” “embezzle[ment],” or “by false pretences.” Mass. Gen. Laws ch. 277, § 41.

2.

Baptista seizes on the comment in Mills that “steal” is now a “term of art” that includes “larceny, embezzlement or obtaining by false pretences.” 764 N.E.2d at 860. He presses that “steal” should be construed that way across Massachusetts law, including in the unarmed assault statute.

If Baptista is right, the use of the term “steal” in his statute of conviction would incorporate larceny, embezzlement, and larceny by false pretenses—situations where a taking occurs with the victim's fraudulently obtained consent. So unarmed assault would criminalize more conduct than a generic federal theft offense.

The Board of Immigration Appeals rejected this argument, reasoning that there is “no indication that [the Mills court's] understanding of the term ‘steal’ was intended to apply outside the context of the larceny statute and the specific purposes for which it was broadly drafted.” J.A. 5. We agree with the Board.

The statutory text doesn't define the word “steal” and the case law interpreting Massachusetts unarmed assault with intent to rob or steal is slim. But we find some guidance in how Massachusetts defines the offense of robbery.

3.

Unarmed assault with intent to rob or steal falls within chapter 265 of the Massachusetts General Laws, which is labeled “Crimes Against the Person.” Other crimes in this chapter include murder, manslaughter, assault and battery, and, most relevant here, armed and unarmed robbery. Mass. Gen. Laws ch. 265, §§ 1, 13, 13A, 15, 17, 19.

The armed and unarmed robbery statutes reference “steal[ing],” just like the unarmed assault offense of which Baptista was convicted. Id. §§ 17, 19(b). And unarmed assault is a lesser included offense of unarmed robbery. See Commonwealth v. Drewnowski, 44 Mass.App.Ct. 687, 694 N.E.2d 1301, 1304–05 (1998).

There's no Massachusetts model jury instruction for unarmed assault, but the instruction for robbery informs our understanding of “steal” for chapter 265 offenses.³ The key language reads:

The second element the Commonwealth must prove beyond a reasonable doubt is that the defendant took the money or other property with the intent to steal it. This means that the Commonwealth must prove that the defendant took and carried away property, against [the alleged victim's] will, with the intent to deprive [the alleged victim] of (his/her) possessions permanently.

Mass. Super. Ct. Crim. Prac. Jury Instr. § 3.16.3 (emphasis added); see also id. § 3.16.1.

The instruction states that the taking of property must be against the victim's will for a defendant to be convicted of robbery. In other words, a person can't commit robbery in Massachusetts through fraud or embezzlement.⁴

Baptista responds that the requirement that the taking be “against the victim's will” refers to the actus reus of the offense, not the intent required to carry it out. On Baptista's reading, the defendant could intend to take property through fraud or embezzlement, but in fact rob using force and violence.

But the instruction itself says that “the intent to steal” property “means ․ that the defendant took and carried away property ․ against [the victim's] will.” Mass. Super. Ct. Crim. Prac. Jury Instr. § 3.16.3 (emphasis added). By its own terms, the instruction forecloses Baptista's argument.

In short, Massachusetts robbery requires that the taking be against the victim's will. We see no reason why the same wouldn't be true for unarmed assault, which is also a chapter 265 “Crime Against the Person.”

4.

We're also persuaded that the unarmed assault statute's requirement of “force and violence” makes the crime impossible to commit with the victim's consent. This element requires, at minimum, “[p]hysical intimidation” that “produce[s] a reasonable apprehension of bodily injury.” Commonwealth v. Gauthier, 21 Mass.App.Ct. 585, 488 N.E.2d 806, 809 (1986). The intimidating act must be intended “to put the victim in fear of immediate bodily harm.” Commonwealth v. Jorritsma, No. 00-P-1153, 2001 WL 1561861, at *1 (Mass. App. Ct. 2001) (unpublished table decision); see also Commonwealth v. Musgrave, 38 Mass.App.Ct. 519, 649 N.E.2d 784, 787 (1995) (explaining that Massachusetts criminal assault requires “proof of intent to cause fear or apprehension”), aff'd, 421 Mass. 610, 659 N.E.2d 284 (1996).

Baptista argues that the “force and violence [requirement] applies only to the assault, not the intended taking.” Reply Br. at 18. We're unconvinced. A person may commit an assault in Massachusetts through either an attempted battery or a threatened battery. Commonwealth v. Porro, 458 Mass. 526, 939 N.E.2d 1157, 1163 (2010). Attempted battery requires the intent to commit a battery. Id. Threatened battery requires “menacing conduct” performed with the intent “to place the victim in fear of an imminent battery.” Id.

In either case, a person must direct the attempted or threatened use of force at the victim to produce the desired result: the turnover of property in the victim's possession. The attempted or threatened use of force constituting the actus reus of unarmed assault would be gratuitous if the victim could consent to the defendant's taking.

The categorical approach requires “a realistic probability, not a theoretical possibility, that the State would apply its statute to conduct that falls outside the generic definition of a crime.” Gonzales, 549 U.S. at 193, 127 S.Ct. 815. Baptista offers no case or hypothetical showing how one could commit unarmed assault with “force and violence” despite the victim's consent to the taking, and we can't imagine one. Even if there's conceptual space between the force and intent elements, as Baptista urges, our job isn't “to apply ‘legal imagination’ to the state offense.” Moncrieffe v. Holder, 569 U.S. 184, 191, 133 S.Ct. 1678, 185 L.Ed.2d 727 (2013).

Our view that the force requirement makes it impossible to commit unarmed assault while intending to take property with the victim's consent finds support in the elements of Massachusetts robbery. That offense, like unarmed assault, “requires that force and violence be used against the victim or that the victim be put in fear.” Commonwealth v. Johnson, 379 Mass. 177, 396 N.E.2d 974, 977 (1979).

Baptista concedes that robbery's force requirement “makes the taking against the [victim's] will.” Pet'r's Br. at 19 (quotation marks omitted); see Commonwealth v. Jones, 362 Mass. 83, 283 N.E.2d 840, 844 (1972) (distinguishing larceny and robbery based on the robbery victim's “aware[ness] of the application of force which relieves him of his property,” which makes robbery “against the victim's will” (cleaned up)). But Baptista rejects the same concession for unarmed assault, without explaining the inconsistency in his position.

We find that the force element of Massachusetts's unarmed assault statute excludes the possibility that a person may be convicted of that offense for a taking committed with the victim's consent.

5.

Our decision tracks the Eleventh Circuit's view of a similar issue. In Kemokai v. U.S. Att'y Gen., 83 F.4th 886 (11th Cir. 2023), the court considered whether the Massachusetts armed robbery statute qualifies as a theft offense under the Immigration and Nationality Act. Applying a categorical analysis, the court rejected the notion that the armed robbery statute's reference to intent to steal meant “that larceny, in all of its common-law forms, constitutes an element of robbery.” Id. at 894.

The Kemokai court cited Massachusetts cases making clear that Massachusetts robbery must occur against the victim's will and with force and violence. Id. at 894–95 (collecting cases). These requirements also apply to the unarmed assault statute, as we've explained.

So we reject Baptista's contention that the unarmed assault statute is categorically overbroad because it allows for prosecution of an act committed with the intent to take property with the victim's fraudulently obtained consent. Rather, we conclude that the completed offense contemplated by the Massachusetts unarmed assault statute qualifies as a theft offense under the Immigration and Nationality Act.

B.

Next, we consider whether Massachusetts unarmed assault with intent to rob or steal categorically matches a generic federal attempt offense. Having concluded that the intent required to violate the statute is the intent to commit a theft offense, we turn to whether the statute requires “an overt act constituting a substantial step towards the commission of [a theft].” Rendon, 764 F.3d at 1084.

1.

To begin, we find unpersuasive the contention pressed by Baptista and amici that an attempt offense under the Immigration and Nationality Act refers only to an offense that a state has labeled as an attempt, such as a general attempt statute proscribing the attempt to commit the completed offenses described in other provisions. Instead, we conclude that the “attempt” referenced in § 101(a)(43)(U) can include a completed offense, the elements of which satisfy the generic federal definition of an attempt.

In applying the categorical approach, our job is a narrow one. We “consider only the elements of the offense of conviction” to determine whether the state offense sweeps in more conduct than the generic federal crime. Castendet-Lewis v. Sessions, 855 F.3d 253, 260 (4th Cir. 2017). And the sole question is whether the minimum conduct that the state statute criminalizes falls within the generic federal crime.

The label a state applies to its crimes—choate or inchoate—is irrelevant. A uniform body of federal law requires definitions that can't be commandeered by magic words. See Taylor v. United States, 495 U.S. 575, 592, 110 S.Ct. 2143, 109 L.Ed.2d 607 (1990) (explaining that a federal sentencing enhancement for predicate burglaries “must have some uniform definition independent of the labels employed by the various States' criminal codes”). We find persuasive a sister circuit's explanation “that Congress did not intend to allow the [federal] definition of ‘attempt’ to vary solely according to the labels applied by the jurisdiction in which an alien was charged.” Sui v. INS, 250 F.3d 105, 114 (2d Cir. 2001); see also United States v. Martinez-Garcia, 268 F.3d 460, 465 (7th Cir. 2001) (adopting Sui's reasoning).

Our rejection of a per se prohibition on completed offenses qualifying as attempts under the Immigration and Nationality Act isn't novel. Several courts, for example, have concluded that the completed offense of entering a locked vehicle with the intent to steal it is an attempted theft offense. Martinez-Garcia, 268 F.3d at 466; Ngaeth v. Mukasey, 545 F.3d 796, 802 (9th Cir. 2008); see Lopez-Elias v. Reno, 209 F.3d 788, 792 n.7 (5th Cir. 2000) (stating in dicta that “burglary of a vehicle with intent to commit theft is tantamount to an offense of attempted theft” (emphasis in original)).

Baptista suggests that breaking into a locked vehicle with the intent to steal is “unique” because of “the specificity of the proscribed conduct.” Pet'r's Br. at 38 n.6. But his suggestion concedes that a completed offense can sometimes be attempted theft. And he offers no reason why other state offenses couldn't likewise be a categorical match for an attempted theft.

Baptista claims that the government's efforts to remove him based on his unarmed assault conviction would “dangerously expand” the Immigration and Nationality Act. Pet'r's Br. at 37. He insists that this leap “creates redundancies in state criminal codes” and “unpredictability in the immigration consequences of criminal convictions.” Id. at 39.

Amicus curiae Legal Aid Justice Center similarly contends that the Board's view “improperly conflates intent with attempt,” which “unduly eas[es] the burden on the Department of Homeland Security ․ to establish removability.” Br. of Amicus Curiae Legal Aid Justice Center at 4 (emphasis in original). Legal Aid Justice Center presses that the Board's position raises a litany of “due process concerns.” Id. at 9. And other amici note that the Board's position would disrupt criminal practice by making it harder for lawyers to counsel their clients about the immigration consequences of pleading guilty to state offenses. See generally Br. of Amici Curiae The American Immigration Lawyers Ass'n et al.

We don't share these concerns. Our conclusion is the natural result of the categorical approach, which requires us to look only at the elements of the state offense. Before today, defense counsel needed to consider the categorical approach to advise their clients on the immigration consequences of criminal convictions, and today's decision does nothing to change that. See, e.g., United States v. Swaby, 855 F.3d 233, 240 (4th Cir. 2017).

In short, we reject the notion that, as a matter of law, offenses defined as complete under state law can't qualify as attempts under the Immigration and Nationality Act.

2.

The final question before us is whether Massachusetts unarmed assault requires “an overt act constituting a substantial step towards the commission of [a theft].” Rendon, 764 F.3d at 1084. Baptista challenges this element on two fronts.

First, he argues that unarmed assault doesn't require proof of an overt act that meets the substantial step threshold. On this point, he relies heavily on Commonwealth v. Delgado, 367 Mass. 432, 326 N.E.2d 716 (1975). In that case, four men committed armed robberies of two convenience stores. During one of the robberies, while a co-defendant held a knife to the store manager, the defendant said, “Hold him or I'm going to shoot him.” Id. at 717. The Delgado court upheld the defendant's assault conviction based on the threat, despite the absence of direct evidence that the defendant had a gun, because the defendant's statement in the context of a robbery could “cause the victim reasonable apprehension with respect to his physical security.” Id. at 719.

According to Baptista, Delgado suggests that words alone aren't necessarily an overt act since verbal threats don't need to “strongly corroborat[e] the defendant's intent, or be capable of actually effecting that intent” to support an assault conviction. Pet'r's Br. at 54. But Delgado says no such thing.

The case makes clear that “words are [generally] not sufficient to constitute an assault,” but distinguishes between “informational words and threatening words.” 326 N.E.2d at 718–19 (quotation omitted). The former category “might take the place of a threatening movement or gesture and complete the assault” by “placing another in reasonable apprehension that force may be used.” Id. at 719 (quotation omitted). Informational words can therefore transform a verbal threat of battery into a completed assault by placing the victim in “reasonable apprehension of bodily injury.” Gauthier, 488 N.E.2d at 809.

In short, because a victim must always be at least objectively afraid of bodily injury even in the threatened battery form of assault, there must be an overt act causing such fear. We therefore reject Baptista's notion that Massachusetts unarmed assault doesn't require proof of an overt act.

Baptista's final challenge is based on his view that Massachusetts law criminalizes assaults after a larceny. On his reading, unarmed assault can't necessarily be a substantial step toward the commission of a larceny because the assault can occur after the defendant abandons his or her intent to steal.

But that's wrong. The cases Baptista cites don't criminalize assaults that take place after a larceny. They show instead that Massachusetts courts affirm assault convictions where the defendant's conduct shows a continuing intent to steal when the assault occurs.

Consider Commonwealth v. Correia, 17 Mass.App.Ct. 233, 457 N.E.2d 648 (1983). There, the defendant stole the victims' handbags from their hotel room while they were sleeping, reentered the room, and then assaulted them. The court found “sufficient evidence for the jury to conclude that the defendant returned to the bedroom with a continuing intent to steal ․” Id. at 651. It rejected the notion that the defendant completed the larceny before the assaults because he returned to the hotel room, suggesting that he intended to take more items. Id.

Likewise, in Commonwealth v. Assad, 19 Mass.App.Ct. 1007, 476 N.E.2d 629 (1985), the court found that the evidence supported the jury's finding that the defendant had broken into the victim's apartment to steal. After the victim in Assad returned to his apartment to find “two strangers,” one of the strangers “followed him, sprayed him with a liquid which [he] took to be mace, and knocked him down.” Id. at 630.

On appeal of the defendant's conviction for assault with a dangerous weapon in a dwelling house with intent to rob, the defendant argued that “there was no evidence of intent to rob.” Id. at 631. The defendant insisted that he “used the spray can to facilitate an escape from the victim's apartment, not to steal from the victim or in his presence by putting him in fear, i.e., with the intent to rob the victim.” Id.

The Assad court reasoned that the jury could have found that “the two men who had broken into an apartment,” “armed ․ with chemical spray devices,” “were there to steal.” Id. As for the link between the intent to rob and the assault, the court noted that the assault “was connected to the objective of stealing property from [the] apartment,” even if it occurred during an escape. Id.

Correia and Assad show that the assault must be linked to a defendant's ongoing aim of robbing or stealing from the victim. So long as the defendant assaults a victim to further that aim, the evidence can support a conviction. That's enough for the assault to constitute a substantial step toward the commission of the theft offense, and so Massachusetts unarmed assault is a categorical match to an attempted theft for purposes of the Immigration and Nationality Act.

IV.

For the reasons given, Baptista's petition for review is

DENIED.

FOOTNOTES

1.   Baptista doesn't challenge that the punishment for a completed theft in Massachusetts is a prison term of at least one year.

2.   If the crime of conviction isn't a categorical match, we consider whether the “[state] statute sets out multiple, alternative elements of a crime, effectively creating several different crimes, and is therefore divisible.” Gomez-Ruotolo v. Garland, 96 F.4th 670, 678 (4th Cir. 2024) (quotation omitted). As explained below, we reject Baptista's argument that his statute of conviction's “intent to rob or steal” element is categorically overbroad as to the “steal” prong. Thus, we need not decide whether the statute is divisible as to “rob or steal.” See Descamps v. United States, 570 U.S. 254, 257, 133 S.Ct. 2276, 186 L.Ed.2d 438 (2013).

3.   The larceny statute the court discussed in Mills falls in chapter 266, which is labeled “Crimes Against Property.” Mass. Gen. Laws ch. 266, § 30.

4.   In applying the categorical approach, we often look to model jury instructions to clarify unclear points of state law. E.g., Pugin v. Garland, 19 F.4th 437, 452 (4th Cir. 2021).

DIAZ, Chief Judge:

Petition denied by published opinion. Chief Judge Diaz wrote the opinion, in which Judge Harris and Judge Berner joined.